From MLY%AI.AI.MIT.EDU at MINTAKA.LCS.MIT.EDU  Fri May 26 20:08:17 1989
From: MLY%AI.AI.MIT.EDU at MINTAKA.LCS.MIT.EDU (Richard Mlynarik)
Date: May 26 89 14:08:17 EDT
Subject: jobdev;chaos tcp
Message-ID: <600669.890526.MLY@AI.AI.MIT.EDU>

The lisp machines won't attempt to use TCP-GATEWAY on MC because
MC doesn't have an internet address.  So what's the problem?

From sra at lcs.mit.edu  Fri May 26 19:52:36 1989
From: sra at lcs.mit.edu (Rob Austein)
Date: May 26 89 13:52:36 EDT
Subject: system full
In-Reply-To: David Chapman's message of Fri, 26 May 89 12:44:56 EDT <600506.890526.ZVONA@AI.AI.MIT.EDU>
Message-ID: <8905261752.AA00896@bigboote.LCS.MIT.EDU>

   Date: Fri, 26 May 89 12:17:38 EDT
   From: David Chapman <ZVONA%AI.AI.MIT.EDU at mintaka.lcs.mit.edu>

   AI was dead in the water, typing Warning:system full, no job slots
   on the console every two seconds.  I crash dumped to crash;sys ful.  I
   don't know whether this is relevant, but it was getting a lot of
   top-level interrupt 200s in TCP and CHAOS jobs not long before it lost
   it. 

   Date: Fri, 26 May 89 12:44:56 EDT
   From: David Chapman <ZVONA%AI.AI.MIT.EDU at mintaka.lcs.mit.edu>

   Uh, actually, maybe the problem was that it was running ITS 1632
   (which RS brought up this morning) -- maybe something else has changed
   incompatibly, like say COMSAT?.  We're running 1633 again now.

Given that ITS.1632 has TCP code which would presumably be somewhat
upset at the lack of an IMP behaving in any known fashion, this is not
terribly surprising.

Yes, COMSAT would also be upset to find that AI was running ITS.1632
again, since it determines what it should do about routing to TCP
hosts by asking the monitor if it supports TCP or not.

Perhaps we should delete/hide JOBDEV;CHAOS TCP on AI and MC?  It's not
like it's doing all those Lisp Machines any good to try getting to the
ARPANET via MC and AI anymore.

From ALAN%AI.AI.MIT.EDU at MINTAKA.LCS.MIT.EDU  Fri May 26 19:30:39 1989
From: ALAN%AI.AI.MIT.EDU at MINTAKA.LCS.MIT.EDU (Alan Bawden)
Date: May 26 89 13:30:39 EDT
Subject: No subject
Message-ID: <600605.890526.ALAN@AI.AI.MIT.EDU>

Right.  There were a bunch of things all going wrong here, most
of them caused by the fact that the hosttable no longer
contains IP addresses for AI and MC.  Certainly that is what
made COMSAT crash, and I'm pretty sure that that is what caused
the system to fill up with dead mail servers.

Rob, you should learn to be more suspicious of situations where
you can't see any reason how something could fail to work.
(The current series of Calvin & Hobbes comes to mind...)

From ZVONA%AI.AI.MIT.EDU at MINTAKA.LCS.MIT.EDU  Fri May 26 18:44:56 1989
From: ZVONA%AI.AI.MIT.EDU at MINTAKA.LCS.MIT.EDU (David Chapman)
Date: May 26 89 12:44:56 EDT
Subject: system full
Message-ID: <600506.890526.ZVONA@AI.AI.MIT.EDU>

Uh, actually, maybe the problem was that it was running ITS 1632
(which RS brought up this morning) -- maybe something else has changed
incompatibly, like say COMSAT?.  We're running 1633 again now.

From ZVONA%AI.AI.MIT.EDU at MINTAKA.LCS.MIT.EDU  Fri May 26 18:17:38 1989
From: ZVONA%AI.AI.MIT.EDU at MINTAKA.LCS.MIT.EDU (David Chapman)
Date: May 26 89 12:17:38 EDT
Subject: No subject
Message-ID: <600479.890526.ZVONA@AI.AI.MIT.EDU>

AI was dead in the water, typing Warning:system full, no job slots
on the console every two seconds.  I crash dumped to crash;sys ful.  I
don't know whether this is relevant, but it was getting a lot of
top-level interrupt 200s in TCP and CHAOS jobs not long before it lost
it. 

From sra at lcs.mit.edu  Tue May  2 23:48:20 1989
From: sra at lcs.mit.edu (Rob Austein)
Date: May 2 89 17:48:20 EDT
Subject: host tables in the brave new world
In-Reply-To: "Patrick G. Sobalvarro"'s message of Tue,  2 May 89 17:33:17 EDT <589618.890502.PGS@AI.AI.MIT.EDU>
Message-ID: <8905022148.AA01196@bigboote.LCS.MIT.EDU>

   Date: Tue,  2 May 89 17:33:17 EDT
   From: "Patrick G. Sobalvarro" <PGS at ai.ai.mit.edu>

   When we switched host table formats a few years ago, I stopped knowing
   how to update them.  Is there an easy way to add AI.MIT.EDU to them, now
   that that's showing up as a return address on mail from the Unixes?

You don't want to know.  Really.  You don't.

The short answer is that everything came to a grinding halt when XX
went away and we are just about back on our feet now.  If all goes
well, automatic table generation and installation should start up
again within a few days.

From PGS at AI.AI.MIT.EDU  Tue May  2 23:33:17 1989
From: PGS at AI.AI.MIT.EDU (Patrick G. Sobalvarro)
Date: Tue,  2 May 89 17:33:17 EDT
Subject: host tables in the brave new world
Message-ID: <589618.890502.PGS@AI.AI.MIT.EDU>

When we switched host table formats a few years ago, I stopped knowing
how to update them.  Is there an easy way to add AI.MIT.EDU to them, now
that that's showing up as a return address on mail from the Unixes?